6.2 The ISO/OSI Reference Model How do you swallow an elephant? A bite at a time. This is the methodology used by ISO in developing a reference model for network communications. The OSI reference model is a guideline for the functionality of networking standards and protocols. Figure 6.1 presents the structure of this reference model. It divides the problem of communicating between systems into individual, bite- sized tasks . These subtasks are represented as layers in the reference model. It then solves each subtask. The stacks in the figure each reside on a separate system. An individual layer in the stack on one system communicates with that same layer on the other system. Communications between the two layers is carried out indirectly through the lower levels of the model. Figure 6.1. ISO/OSI reference model.  For example, an application in one system wishes to communicate with an application in another. To carry out this communication, the application communicates to the presentation layer that in turn communicates with the session layer. This process repeats until the message passes between the two systems over the physical layer. The message then makes its way up the stack to the presentation layer on the destination system. In Chapter 5, we showed how Internet-enabled applications interact with the browser. The browser in turn invokes the operating system. We can see how the communication proceeds down from the application layer of one system, across the network, and back up the stack to the destination application. In reviewing the model, we should note that this is a reference model ; it is not a specification for a particular protocol or implementation. We use this reference model as a way to understand and organize other protocols and standards. To understand this model more fully, let's review the function of each layer. -
Physical Defines the actual hardware connection between systems. Physical specifications include specification of the connector type. Is it a USB port, RS-232, or RJ-49? These are all different types of connectors. The physical layer also describes the electrical specification between systems. Is the network analog or digital? It describes something about that actual physical connection between systems. -
Data Link Organizes the stream of bits it receives from the physical layer into frames . A frame is composed of a header and trailer that contain control information. The data link layer, in addition to detecting errors in the data, controls the flow of information between the systems. -
Network This layer's task is to move information across the network. The network layer examines the network layer address of a packet of information and passes it on to the next subnetwork within the Internet via a router. -
Transport This layer provides reliable delivery of data across the network. Note that there is a significant difference between ensuring reliable delivery and guaranteeing delivery of data. The transport layer will detect when a message is not acknowledged as having been received by the destination system. It then informs the session layer and allows the upper layers to take whatever action may be necessary. -
Session This layer controls the flow of information between the systems. It manages the dialog between the systems that are communicating with one another. -
Presentation The presentation layer converts the bits into something meaningful. It is at the presentation layer that the data is compressed or encrypted. -
Application The final layer, the application layer, manages the communication between the user or application and the network. The application layer includes virtual terminals and mail transfer services. As information proceeds down through the ISO/OSI reference model, each layer may add a header and in some cases a trailer to the data. This additional information is used to communicate to the corresponding layer on the destination system. For example, the data link layer, which is responsible for detecting errors in the data, might add parity bits to the message as it passes from the network layer to the physical layer. The data link layer on the destination system will take the data it receives from the physical layer and reconstruct the header built on the originating system. It is then able to use this header information to detect errors and possibly correct them. Let's look at one more example to demonstrate this point. We noted that a cookie is a block of text. An application may generate a message containing a cookie. This message is composed of the cookie itself and some control information contained in the message header. The presentation layer encrypts the entire message, including the header created by the application layer. Once the message is encrypted, the presentation layer adds its own headers and trailer. The presentation layer on the destination system uses this to unscrambled the encrypted message. The session level receives the message from the presentation layer, complete with the presentation layer header, and passes it on to the transport layer. The transport layer adds its own set of headers and trailers to create a segment. The segment is passed to the network layer, which constructs a datagram by adding its own headers and trailer. The data link builds a frame, which is passed to the physical layer, where the frame is passed to the destination system as a stream of bits. The destination system performs the same functions in reverse. The stream of bits is used to construct a frame. The control information in the header and trailer of the frame tells the data link layer how to process the data contained in the frame. The cookie proceeds up the ISO/OSI stack, with each layer peeling off the associated header. Finally, the cookie arrives at the application layer, where it is received by the destination application. We will discuss the concept of packets again later in this chapter. |
